Of all multifactorial child psychiatric disorders, autism is the most strongly genetically influenced. The inherited liability is not restricted to the full clinical syndrome of autism but encompasses a range of behavioral and cognitive characteristics. Family studies have failed to distinguish whether the genetic liability to autism falls on a continuous spectrum of severity or whether it may be better subdivided into distinct categorical subtypes. From a genetic standpoint this raises the question, whether different aspects of the autism phenotype are influenced by different loci. The proposed study will be the largest, population-based study of twins with autism. With this unique sample we will be able to unravel the relationship between genes and environment as it pertains to the cognitive impairments and the clinical symptoms seen in children with autism spectrum disorders. We will assess twin pairs using (a) clinical measures of autistic-like behaviors (b) more specific measures for repetitive and stereotyped behaviors, (c) measures of general cognitive ability d) measures of more specific neurocognitive abilities. This will allow us to address several fundamental questions: (1) what is the heritability of autism (2) what is the contribution of genetic factors to variation in symptom dimensions? (3) is there a continuum between the quantitative neurocognitive traits and clinical disorder? (4) what proportion of the variance in the neurocognitive traits is accounted for by genetic and non-genetic factors?